Cytosolic action of phytochelatin synthase

Glutathionylation of compounds is an important reaction in the detoxification of electrophilic xenobiotics and in the biosynthesis of endogenous molecules. The glutathione conjugates (GS-conjugates) are further processed by peptidic cleavage reactions. In animals and plants, γ -glutamyl transpeptidases initiate the turnover by removal of the glutamic acid residue from the conjugate. Plants have a second route leading to the formation of γ -glutamylcysteinyl- ( γ -GluCys-) conjugates. Phytochelatin synthase (PCS) is well known to mediate the synthesis of heavy metal binding phytochelatins. In addition, the enzyme is also able to catabolize GS-conjugates to the γ -GluCys-derivative. In this study, we addressed the cellular compartmentalization of PCS and its role in the plant specific γ -GluCys-conjugate pathway in Arabidopsis. Localization studies of both Arabidopsis PCS revealed a ubiquitous presence of AtPCS1 in Arabidopsis seedlings, while AtPCS2 was only detected in the root tip. A functional AtPCS1:eGFP fusion protein was localized to the cytosolic compartment. Consistently, inhibition of the vacuolar import of GS-bimane conjugate via azide treatment resulted in both a strong accumulation of γ -GluCys-bimane and in a massive increase of the cellular Cys- to GS-bimane ratio, which was not observed in PCS-deficient lines. The findings support a cytosolic action of PCS. Analysis of a triple mutant deficient in both Arabidopsis PCS and vacuolar γ -glutamyl transpeptidase GGT4 is consistent with earlier observations of an efficient sequestration of GS-conjugates into the vacuole and the requirement of GGT4 for their turnover. Hence, PCS contributes specifically to the cytosolic turnover of GS-conjugates and AtPCS1 plays the prominent role. We discuss a potential function of PCS in the cytosolic turnover of GS-conjugates.


INTRODUCTION
Plants are sessile organisms that cannot evade unfavorable habitats. They have therefore developed effective means to overcome stress conditions such as herbivore and pathogen attack, or abiotic challenges encountered by drought, light stress, high levels of metal ions, and exposure to herbicides. Under such conditions, the tripeptide glutathione (GSH) plays a number of pivotal roles in plants (Noctor et al., 2002;Schützendübel and Polle, 2002). GSH is present in the millimolar range and constitutes up to 90% of total non-protein sulfur (Leustek et al., 2000). The tripeptide is an important component of the cellular redox system (Foyer et al., 2009) and is connected to the regulation of stress-responsive gene expression (Ball et al., 2004;Mullineaux and Rausch, 2005). The redox-coupled formation of mixed disulfides among plant proteins is considered to be an important mechanism for posttranslational regulation (Dixon et al., 2005). The conjugation of electrophilic molecules to GSH is catalyzed by the action of glutathione S-transferases (GSTs). In plants, GSTs are encoded by a large gene family with approximately 50 members in Arabidopsis and rice (Soranzo et al., 2004;Edwards and Dixon, 2005), highlighting the importance of GSconjugate formation for the metabolism of endogenous compounds and the detoxification of noxious compounds such as herbicides. GS-conjugates are predominantly generated in the cytosol with minor GST activities in the nucleus, chloroplast, and mitochondrion (Dixon et al., 2002;Dixon and Edwards, 2009).
The rapid vacuolar sequestration of GS-conjugates is mediated by ATP-binding cassette (ABC) transporters (Martinoia et al., 1993;Edwards et al., 2000). The ABC multidrug resistance-associated proteins (MRPs) AtMRP1, AtMRP2 and presumably AtMRP3 (Rea, 2007) have been characterized as GS-conjugate transporters. In the vacuole, degradation of GS-conjugates is initiated by the γ -glutamyl transpeptidase AtGGT4 ( Fig. 1) yielding CysGlyderivatives (Grzam et al., 2007;Ohkama-Ohtsu et al., 2007b). The cleavage of the unique isopeptidic bond of GS-conjugates by GGTs, which corresponds to the first step in the catabolism of GS-conjugates, and the subsequent enzymatic reactions are well characterized in animal cells (Martin et al., 2007). However, the enzymatic activity for the degradation of the resulting CysGly dipeptide conjugates to the Cys-conjugates ( Fig. 1) has not been unequivocally identified in plants. In yeast, the first degradation step is mediated by a single GGT, which is also required for extracellular secretion of Cys-conjugates (Wünschmann et al., 2010). Plants possibly recruit the GGTs in a similar manner (Martin et al., 2007) or alternatively, γ -glutamyl cyclotransferase and 5-oxoprolinase, which can convert the glutamyl moiety of While AtGGT4 plays a role in GS-conjugate catabolism in the vacuole, AtGGT1 and AtGGT2 have an active site in the apoplast (Ohkama-Ohtsu et al., 2007a) and could be involved in the secretion of Cys-conjugates coupled with the cleavage of the glutamyl moiety from γ -GluCys-conjugates (Fig. 1).
A second mode of GS-conjugate catabolism is initiated by the removal of the carboxyterminal Gly to the corresponding γ -GluCys-product (Fig. 1). This catabolic step is common for GSconjugate degradation in plants (Lamoureux andRusness, 1986, 1993). A vacuolar carboxypeptidase was postulated to catalyze the deglycination of GS-conjugated molecules of the herbicide alachlor in barley (Wolf et al., 1996). The removal of the carboxyterminal Gly from GS-conjugates is efficiently catalyzed by phytochelatin synthase (PCS, EC 2.3.2.15) (Beck et al., 2003;Blum et al., 2007). PCS is a specific γ -GluCys dipeptidyl transpeptidase (Grill et al., 1989;Vatamaniuk et al., 2004), known to generate the heavy metal chelating phytochelatins with the general structure (γ-GluCys) n -Gly, n = 2-11 from GSH (Grill et al., 1985). Arabidopsis expresses two PCS proteins, AtPCS1 and AtPCS2 (Cazale and Clemens, 2001). Studies with the model xenobiotic monochlorobimane (MCB) suggest that AtPCS1 rather than AtPCS2 is the dominant player, as it has been shown that functional disruption of AtPCS1 but not of AtPCS2 leads to a decrease in the turnover of the fluorescent GS-bimane conjugate to γ -GluCys-bimane in Arabidopsis (Blum et al., 2007). Assessment of PCS contribution to GS-conjugate catabolism has been hampered by the existence of a second GGT-initiated pathway.
Both, the glucosinolate catabolism and the immune response were impaired in the AtPCS1deficient mutant. Collectively, the data support the existence of a cytosolic activity responsible for γ -GluCys-conjugate formation in plants, which might be provided by PCS.
In this contribution, we localized AtPCS1 to the cytosolic compartment and addressed the extent of PCS contribution to GS-conjugate catabolism. Inhibition of GS-conjugate transport into the vacuole resulted in a strong accumulation of γ -GluCys-bimane. The accumulation of

Blocking vacuolar transport by azide
GS-conjugates are transported to the vacuole, where they are degraded. The catabolism of GS-conjugates may also occur in the cytoplasm. As PCS is known to catabolize GSconjugates to the γ -GluCys-derivative, it is a prime candidate for mediating an alternative metabolic pathway for GS conjugates in the cytosol. MCB is a model xenobiotic that has successfully been used to track the catabolism of GSbimane conjugates in planta (Grzam et al., 2006;Blum et al., 2007;Grzam et al., 2007).
MCB is typically conjugated to GSH in the cytosol and subsequently sequestered in the vacuole. In order to evaluate the consequence of inhibiting the vacuolar sequestration of MCB, we treated cells with azide, a potent inhibitor of F-ATPases (Bowler et al., 2006) and of ABC transporters (Dallas et al., 2003;Jha et al., 2003). Confocal images of MCB-challenged cell suspension cultures of Arabidopsis show that the rapid accumulation of the fluorescent signal in the vacuole was prevented in the presence of azide ( Fig. 2A), as has been reported in earlier studies (Meyer and Fricker, 2002;Grzam et al., 2006). We used feeding experiments with MCB in the presence of azide to assess whether blocking the uptake of bimane-conjugates into vacuoles results in a conversion of the GS-conjugate in the cytoplasm. To analyze the effect of azide on GS-bimane turnover, the optimal azide concentration was established. In the concentration range tested from 1 µM to 1 mM azide, an enhanced turnover of GS-bimane to γ -GluCys-bimane could be observed with a maximum at 1 mM azide yielding 8.5% γ -GluCys-bimane of total bimane labeled compounds compared to 0.5% γ -GluCys-bimane without azide 30 minutes after MCB challenge (Fig. 2B,C).
Interestingly, the conversion level of the GS-conjugate to the Cys-adduct was higher in the presence of 0.3 mM azide compared to non-inhibited cells (Fig. 2B). In the presence of 0.3 mM azide, Cys-bimane accounted for 36% of bimane-labelled compounds, while total recovery dropped to approximately 40% in comparison to untreated cells. The lower recovery of inhibited cells may reflect stimulated secretion of bimane derivatives. Vital staining of cells with fluorescein diacetate provided no evidence for the loss of cellular integrity by azide treatment (Fig. S1 A). The determination of cellular GSH contents in the presence of various azide concentrations did not show significantly altered GSH/GSSG levels ( Fig. S1 B). Subsequent time course experiments revealed a maximum of γ -GluCys-bimane levels 30 min after a pulse of MCB challenge (Fig. 3A, B). Arabidopsis cells exposed to azide yielded a 23-fold elevated γ -GluCys-bimane level compared to the control (13.9 and 0.6% of total bimane derivatives, respectively). Azide treatment (1 mM) did not prevent the degradation of GS-bimane to Cys-bimane in wild type (Fig. 2B) and PCS-deficient cells (Fig. 3C).
The data are consistent with an efficient PCS-dependent generation of the cleavage product in the cytosol. The GS-to Cys-conjugate conversion is slowed down in Δ PCS compared to the wild type in the presence of azide (Fig. 3C), however, the conversion in

Complementation of PCS1-deficiency by AtPCS1-eGFP
AtPCS1 initiates the GS-conjugate catabolic pathway that presumably occurs in the cytosol. lines the concentration increased to 4.2, 5.8 and 8.8%, respectively. Taken together, the findings document a successful complementation of Δ PCS1 by AtPCS1:eGFP.

Localization of PCS proteins in Arabidopsis
PCS might be a soluble or a membrane-associated protein.
To address the issue of intracellular localization of AtPCS1, cell-free extracts of AtPCS1:eGFP expressing seedlings were separated into a soluble protein fraction and a microsomal-enriched fraction by differential centrifugation. Comparable protein levels from both fractions were analyzed for the presence of the tagged AtPCS1 protein by immunodetection (Fig. 7). The AtPCS1 fusion protein was primarily detected in the soluble protein fraction at the molecular mass of approximately 80 kDa, which corresponds to the sum of 55 kDa for AtPCS1 and 26 kDa for eGFP. The microsomal fraction yielded low levels of AtPCS1:eGFP possibly due to a weak interaction of the fusion protein with membranes, or to a contamination of the fraction with soluble proteins such as observed for the GFP control and indicated by the presence of rubisco (Fig. 7).
Analyses of PCS:eGFP expression by confocal imaging revealed a major difference between AtPCS1 and AtPCS2 expressing lines. AtPCS1:eGFP was readily detectable throughout the seedling ( Fig. 8 A-D). In contrast, the AtPCS2:eGFP signal was consistently low in more than 20 independent transgenic lines analyzed and the fusion protein was only detected in the root tip, preferentially in the tunica (Fig. 8 E-G). The low expression levels of AtPCS2, however, might have prevented the detection of the fusion protein in other parts of the seedling and also made it difficult to assess the intracellular localization of AtPCS2, other than to say that it resembled that of AtPCS1 described below (Fig. 9).
AtPCS1 was predominantly expressed in the epidermal layers of the shoot and root including root hairs, but weakly expressed in guard cells (Fig. 8, 9A). In epidermal cells, AtPCS1:eGFP was found in the cytosol, where it could be seen in the cytoplasmic strands of the highly vacuolated cells (Fig. 9 A). Comparison of the AtPCS1:eGFP lines with marker lines in which GFP is targeted to specific organelles (Cutler et al., 2000) pointed to a similarity to soluble GFP, which is found in the cytosol (Fig. 9 B). The only difference observed was that whereas soluble GFP can be clearly detected in the nucleus of the control line, the AtPCS1:eGFP fusion was not found in the nucleus. In some of the primary transformants, AtPCS1:eGFP labeled diffuse reticulate structures in addition to the cytosol. The analysis of meristematic, largely avacuolate cells of the root tip supported a cytosolic localization of AtPCS1 ( Fig. 9

DISCUSSION
GSH is of pivotal importance for metal ion homeostasis, detoxification of xenobiotics, and the metabolism of sulfur compounds. PCS is involved in these processes by generating the metal binding PCs from GSH and catabolizing GS-conjugates. In this contribution, we localize AtPCS1 to the cytosol and provide evidence for a cytosolic action of the enzyme.
While PCS is responsible for cytosolic Taken together, PCS is the prominent activity for initiating the cytosolic catabolism of GSconjugates (Fig. 1). Simultaneously, PCS functions in the generation of the metal-binding PCs. Both enzymatic activities are stimulated by heavy-metal ions (Beck et al., 2003;Blum et al., 2007) and the findings raise the question, how the two different reactions are regulated.
PCs are generated by the transpeptidation reaction of PCS in which the products, PCs, are scavenging the PCS-activating heavy metal ions. PCS is able to degrade metal-free PCs into

Stress exposure and labeling with monochlorobimane
Arabidopsis seedlings were raised on agar solidified MS plates and used for stress exposure and MCB labeling experiments five days after germination. For the analysis of PCs, seedlings were transferred to 20 ml liquid MS medium containing 50 µM Cd(NO 3 ) 2 and incubated for 48 h prior to extraction and HPLC analysis. Cadmium sensitivity was examined by determination of root growth within three days after transfer of the seedlings on MS plates containing up to 100 µM Cd(NO 3 ) 2 . For the analysis of bimane conjugates, seedlings were vacuum-infiltrated for 10 min in MS medium containing 5 µM MCB.
Arabidopsis protoplasts were prepared from 3 week-old rosette leaves as described (Himmelbach et al., 2002) and approximately 5x10 4 protoplasts (0.5 ml) were incubated with 5 µM MCB for 4 h prior to extraction of bimane derivatives (Blum et al., 2007). For experiments with azide (NaN 3 ), 9.6 g (fresh weight) of Arabidopsis cell culture material was transferred to 100 ml fresh LS-medium (Linsmaier and Skoog, 1965) and allowed to recover for 24h. The cells were preincubated with 1 mM azide for 5 min, subsequently pulsed with 5 µM MCB for 1 min, washed (25 ml), and then incubated in MCB-free medium in the presence or absence of azide as indicated. Aliquots of 10 ml cell suspension were removed at different time points. Harvested plant material (0.3 g fresh weight) was used for extraction of bimane derivatives and the remaining material was used for the determination of dry weight. For analysis of the GSH and GSSG content, 30 mg of cell material was transferred to 0.2 ml 20 mM Tris-Cl buffer, pH 8, and homogenized by sonification for 15 min. After centrifugation, the cell free supernatant was incubated (1h, 37 °C) with or without 1 mM DTT.
The samples were diluted 1:10 with the Tris-Cl buffer (total volume 50 µl) and incubated with
Briefly, seedlings (0.15 g fresh weight) were frozen in liquid nitrogen, ground to powder and the material incubated with extraction buffer (80 mM glycine, 20 mM NaCl, 10mM EDTA, pH 3). Protoplasts (5x10 4 ) were directly lysed in 0.4 ml of extraction buffer. Precipitated proteins were removed by centrifugation (10 min, 16,000 x g) and the deproteinated extract was subjected to HPLC analysis (Blum et al., 2007). PCs were analyzed by online derivatisation of sulphhydryl groups with Ellman reagent (Grill et al., 1987).

Confocal analysis
For eGFP analysis in transgenic plants, 5 day-old seedlings were examined using a confocal laser scanning microscope (Fluoview FV1000; Olympus, Hamburg, Germany http://www.olympus-global.com). A 40X water immersion 0.9 numerical aperture objective (Olympus) was used, and scanning was carried out with two-fold or without electronic magnification. Images were acquired and processed with the Fluoview 1000 acquisition software. Three dimensional projections of Z stacks of GFP expression in seedlings were generated with the Imaris 6.   γ -GluCys-bimane (■) amounts (D). Cells were exposed to 1 mM azide.
The initial amounts of GS-bimane and bimane derivatives were set to 100% at the onset of the experiment. The total levels were 448, 307, and 207 nmol/g dw bimane derivatives for wt (-and + azide), and for Δ PCS (+ azide), respectively. Data points represent means of duplicates (deviation < 10% of value) and two independent repetitions yielded similar results.
Incubation conditions as mentioned in Fig. 2.   A, Cadmium induced PC formation. Arabidopsis seedlings were exposed to 50 µM Cd 2+ for 24 hours prior to PC analysis. The level of thiol is given for PC 2 (black columns), PC 3 (grey columns), PC 4 (white columns). The error bars indicate SD, n = 3.
B, Root growth assay for Cd 2+ sensitivity. Root growth of 4-day-old seedlings of wild type (□), Δ PCS1 (○), and two independent, non-segregating transgenic Δ PCS1 lines transformed with AtPCS1:eGFP, L1 (■) and L7 (•), was determined after transfer of the seedlings onto solidified nutritional medium containing Cd 2+ as indicated. The values of root extension within 3 days is presented for n = 13 and SD are shown. Analysis of the complemented line L6 yielded comparable results as shown for L1 and L7.
C and D, Analysis of GS-bimane turnover in wild type, Δ PCS1, and three independent transgenic Δ PCS1 lines (L1, L6, L7). Leaf protoplasts (C) and seedlings (D) of these lines were exposed to 5 µM MCB for 4h prior to analysis of fluorescent bimane conjugates. The seedlings were also challenged with 10 µM Cd 2+ . Values of triplicates with SD are given as percentage of the sum of bimane derivatives analyzed (84,2 nmol/g fw bimane derivatives ± 12%).